> ## Documentation Index > Fetch the complete documentation index at: https://docs.erdo.ai/llms.txt > Use this file to discover all available pages before exploring further. # Declares work to run on a schedule: an ordered list of named steps, each an > action to invoke or a script to run, bounded by the provider objects a workstream governs. The declaration replaces whatever was previously declared under the same name; enabled false removes it. REST only, with no MCP write tool, and the asymmetry is the point. The scope a run acts on is read from the workstream at run time, so an automation that could declare its own scheduled work could hand itself a wider one than it was given — which is exactly what moving scope out of prose was for. ## OpenAPI ````yaml /api/openapi.json put /v1/scheduled-actions openapi: 3.0.0 info: title: Erdo API description: >- Erdo's REST API: query and write datasets, run agents, manage threads, integrations, pages, evals, workstreams, and experiments. Authenticate with a Bearer API key (erdo_api_...) or scoped token (erdo_token_...). version: '2026-07-28' servers: - url: https://api.erdo.ai description: Production security: - bearerAuth: [] paths: /v1/scheduled-actions: put: summary: > Declares work to run on a schedule: an ordered list of named steps, each an description: > action to invoke or a script to run, bounded by the provider objects a workstream governs. The declaration replaces whatever was previously declared under the same name; enabled false removes it. REST only, with no MCP write tool, and the asymmetry is the point. The scope a run acts on is read from the workstream at run time, so an automation that could declare its own scheduled work could hand itself a wider one than it was given — which is exactly what moving scope out of prose was for. operationId: PUT:mcp.SetScheduledActionAPI requestBody: content: application/json: schema: properties: destination: $ref: '#/components/schemas/mcp.ScheduledActionDestination' enabled: title: > enabled defaults to true. Passing false removes the declaration. type: boolean name: description: > run several without either replacing the other. Lowercase letters, digits, hyphens and underscores. title: > name identifies this declaration within the organization, so an org can type: string schedule: description: | at eight. title: > schedule is a five-field cron expression, e.g. "0 8 * * *" for every day type: string steps: description: | under its name for the steps that follow. items: $ref: '#/components/schemas/mcp.ScheduledStep' title: > steps is ordered, and every one of them runs. Each step's result is kept type: array timezone: description: > "America/New\_York". Defaults to UTC. It matters for anything comparing against a provider's own day boundary: an ad account reports yesterday in the account's timezone, and a run at 08:00 UTC is the previous afternoon in Miami. title: | timezone is the IANA zone the schedule is read in, e.g. type: string workstream: description: > bound what this run may act on. The refs are read at run time and handed to the steps as {{scope.ids}} and {{scope.refs}}; nothing is copied here, so publishing a new set takes effect on the next run with no edit to this declaration. A workstream governing nothing yet means a run that does nothing and says so — never a run with no filter. Omit this only for a declaration that genuinely acts on no provider objects. title: > workstream is the slug of the workstream whose declared external refs type: string type: object responses: '200': content: application/json: schema: properties: destination: $ref: '#/components/schemas/mcp.ScheduledActionDestination' enabled: type: boolean job_id: description: | history can be inspected through the automation surfaces. title: > job_id names the automation implementing this declaration, so its run type: string name: type: string schedule: type: string steps: items: $ref: '#/components/schemas/mcp.ScheduledStep' type: array timezone: type: string workstream: type: string type: object description: Success response default: $ref: '#/components/responses/APIError' components: schemas: mcp.ScheduledActionDestination: description: | declaration whose steps act rather than report. properties: columns: additionalProperties: type: string description: > {{scope...}}, {{steps.\...}}, {{result...}} for the last step's value, {{item}} when the last step returned a list, and {{now}}. A run whose final step returned a list writes one row per element, which is how a set of findings becomes a set of rows. Anything else writes one row. title: | columns maps each destination column to a template, resolved against type: object dataset: title: | dataset is where result rows are written. type: string key_column: description: > the same conclusion twice replaces it rather than appending a rival copy. A row whose key renders empty is not written. Omit it to append. title: > key_column makes the write an upsert on that column, so a run that reaches type: string title: > Where to record what a run produced. Omitting a destination is normal for a type: object mcp.ScheduledStep: description: > (app and key) OR a script to run, never both. Unlike a dataset row action, every step runs, in the order declared, and each one's result is kept under its name for the steps after it. That is because these steps are the reads a judgement is made from: a step skipped because an earlier one succeeded would remove a threshold's only source without saying so. properties: app: description: > connected integration such as "google\_ads". Omit it for a script step. title: | app names the app an action belongs to — "erdo", or the slug of a type: string for_each: description: > {{placeholder}} resolving to one — {{scope.refs}} runs it once per governed object. The action is invoked once per element, with {{item}} and {{item\_index}} available in its input, and the step's result is the list of what came back. Every invocation counts against the run's budget. title: | for_each fans the action out over a list: a JSON array, or one type: object input: description: > may contain {{placeholder}} references resolved against the run: {{scope.ids}} for the provider ids this run may act on, {{scope.refs}} for the full objects, {{steps.\}} for an earlier step's result, and {{now}}. Everything else passes through untouched, so a numeric setting stays a number. title: > input carries the action's own parameters, as a JSON object. String values type: object key: description: | a script step. title: > key names the action within that app, e.g. "query_google_ads". Omit it for type: string name: description: > \`steps.\\`. Required, unique within the declaration, and a JavaScript identifier: lowercase letter first, then letters, digits and underscores. title: | name is how the steps after this one refer to what it returned, as type: string result_path: description: > Many actions wrap their answer in an envelope that echoes the request, and a path selects the part worth keeping. Defaults to the whole result. title: > result_path is a dotted path into the action's result, e.g. "results". type: string script: description: > body receiving (scope, steps, run): the objects this run may act on, the results of the steps before it addressed by name, and {now, budget, invocations} describing the run. What it returns becomes this step's result. It may also return a \`note\` explaining the decision, which is reported on the run result. A script shapes data: it cannot invoke actions or read datasets, because a declaration has to say what it calls and where it writes. Every \`steps.\\` a script mentions must name a step declared before it. That is checked when the declaration is saved, so removing a read that a threshold depends on fails here rather than evaluating to nothing every morning. title: > script is JavaScript run instead of invoking an action. It is a function type: string title: > One step of a scheduled declaration. A step is EITHER an action to invoke type: object responses: APIError: content: application/json: schema: externalDocs: url: https://pkg.go.dev/encore.dev/beta/errs#Error properties: code: description: Error code example: not_found externalDocs: url: https://pkg.go.dev/encore.dev/beta/errs#ErrCode type: string details: description: Error details type: object message: description: Error message type: string title: APIError type: object description: Error response securitySchemes: bearerAuth: type: http scheme: bearer description: An Erdo API key (erdo_api_...) or scoped token (erdo_token_...). ````